By Kyle Billings from Berklee College of Music's Music Business Journal.
Information is changing the music industry; not only in the ways we consume and discover music, but as well in the ways companies are arming themselves for competition. Streaming services see gold in the figures that fuel recommendation algorithms, and now content companies, driven too by the alluring growth of digital tech, are proudly announcing their latest partnerships with leading software-as-a-service companies and the terabytes of social data multiplying on their servers.
The future of music discovery—and potentially, of artist development—is now in the numbers.
Algorithms are the most influential development in music discovery since the mixtape. Spotify, iTunes, Last.fm, Pandora, Amazon, and YouTube are a fraction of the innumerable services running numbers to build playlists and suggest music to fit moods and tastes. Active listeners can explore similar artists and delve deep into niche genres, while passive listeners can launch endless playlists with a single click. As competition spreads the streaming industry thin, robust recommendations are a desired commodity.
The Echo Nest, an MIT startup from Cambridge, MA, has become the preeminent provider of musical data since its founding in 2005. Empowered by both public grants and venture capital funding, the company’s computer scientists comb the Internet, building arrays of tempo, key, acoustic features, and other aural imprints. As of February of 2014, The Echo Nest boasted nearly 1.2 trillion data points about the songs in its library, yielding the most comprehensive “musical brain” of its kind.
The Echo Nest was originally founded to act as a reference for hundreds of companies, apps, and websites, giving both nascent and established developers the intelligence to build dynamic musical experiences. It has triumphed through multiple rounds of financing, raising a total of $25.6M. Early in March, moreover, the Swedish streaming platform Spotify purchased “The Nest” for between $100M to $125M—the majority being in equity1.
The Echo Nest thrived on neutrality; the “musical brain” grew and gained support as a central database for music intelligence. Spotify’s acquisition will undeniably affect its reputation as the go-to music data API. While many of The Echo Nest’s clients have yet to publically react, the streaming service Rdio has announced the end of its relationship with the company, opting to work more closely with Rovi, an alternate data provider2. Several hopefuls are now eyeing the vacancy including Musikki and OpenAura. Joao Afonso, CEO and co-founder of Musikki, assures that “Spotify competitors won’t have to wait long for a complete solution3.”
Now, with the most advanced database of musical knowledge at its fingertips, and a newly established credit line of $200M, Spotify is likely approaching an IPO. Three years after its US launch, going public is a major milestone.
Data and Content Creation
Can the algorithms that power music discovery on streaming services be adjusted to spot undiscovered hits? In the same way that Spotify searches for the perfect song to energize athletes, can equations identify artists at the earliest stages of their potential? The question is keeping top minds in the music industry awake.
At South by Southwest, a computer called Watson applied inductive reasoning to solve what is now termed a ‘computational creativity’ problem. By combing through a database of recipes, flavors, and meal reactions, Watson was able to create foods that excited people. Similarly, with an encyclopedia of genres, instruments, lyrics, social reactions, and locations, the next world-renowned act could be the output of an ingenious collaborative filtering program.
300 Entertainment Takes to Twitter
In his keynote speech during 2014 MIDEM conference in Cannes, France, Lyor Cohen—former executive of Def Jam and Warner Music—announced a partnership between his new company, 300 Entertainment, and the newly public social media network, Twitter. The agreement allows 300 to analyze Twitter’s data to inform the production of software tools, and to identify trends that may optimize the A&R process.
Musicians and Twitter have long been linked. In April of 2013, the relationship yielded Twitter #Music, a stand-alone music discovery application with strong intentions but an unfortunate fate (there were few adopters). With its latest partnership, Twitter aims to make an impact where #Music fell short. Twitter’s head of music, Bob “Moz” Moczydlowsky hopes to “[build]… something that everybody in the music business can benefit from” by leveraging a dedicated music data set to create advanced A&R and targeted marketing tools that will teach the industry how to “best invest in artists or… direct their marketing campaigns.”
As Twitter gains a redemptive shot in the music industry, 300 Entertainment is looking to access an ideal data set. “The noise of the Internet has drowned out the voices [that matter],” Cohen believes, “wouldn’t it be great if technology today moved past that noise instead of trying to understand it?” In his perspective, the searing individualism that once marked music’s most influential fans has been extinguished by the mass democracy of YouTube views and Facebook likes. If the industry considered all 1s and 0s equally, music would screech to a halt of stylistic stagnancy.
Cohen is positioning 300 to identify the few; he is searching for the same early adopters that snuck out at night to see shows during his earlier years as a record executive. Could there be, for instance, someone in Chicago who, when she tweets, makes a meaningful impact on the exposure of an artist? If so, that is what Cohen would be looking for. He is not interested in what he calls the ‘massive-passive’ audience of today.
Warner Music Links Up With Shazam
Lyor Cohen is not alone in his quest. Hot off the tails of 300’s announced partnership with Twitter, Warner Music linked up with Shazam—a popular mobile application that listens to and identifies music playing nearby its users. Rob Wiesenthal of Warner says that “by partnering with Shazam, a brand which is synonymous with music discovery for fans all around the world, we have forged a potent proposition: the first crowd-sourced, big data record label.”
In any given minute, Shazam is processing over 10,000 songs. Their swelling database, used as a tool by Warner’s tried-and-tested A&R team, will be leveraged to catch unsigned artists right as their popularity begins to surge. Few listeners need to scan hit songs; Shazam is designed as the go-to app for music lovers as they hear something new that catches their ear. Warner is rightly excited by this fact.
“There’s so much information that we’ve never had before as an industry, and Shazam is at the forefront of that,” Wiesenthal said. As Shazam’s 88 million users scan songs they like, a geographical map of music is formed, defining not only what music is on the rise, but also when and where these songs are heard.
Like Twitter, Shazam stands to benefit from its music industry relationship. The app aims to improve user experiences by integrating artist promotion directly into the software, not unlike Bruno Mar’s performance at the Super Bowl or Wiz Khalifa’s release of “We Dem Boyz.” Shazam CEO Rich Riley envisions the application as “the place you go for lyrics, the place you go to see video, the place you go to engage around a particular artist.4”
Jim Lucchese, CEO of the Echo Nest, sums up the state of the music industry as follows: “the massive amount of data that’s available is incredibly exciting, [and the] reality is that there is a scarcity of people out there who really know how to make sense of it.” After a decade or more of missed opportunities, executives have at last opened their arms to the data scientists that may discover the next world-changing artist.
But will data convey the competitive advantage that its advocates intend? Social media can narrate an industry, but can tweets and scans replace the ears of traditional A&R scouts? Both 300 Entertainment and Warner are led by time-tested musical minds, integrating one-part computers to four-parts instinct. At the same time, pioneering experiments in computational creativity are setting the stage for a future of experiments.
Moore’s law holds that computing power is growing exponentially. Fifteen years ago, iPods could not play videos, video blogs weren’t yet called “vlogs,” and laptop computers were still a tentative technology. It is true that programs are unlikely to override the music industry we all know and love anytime soon. For instance, in Lyor Cohen’s world the industry mogul is still the gatekeeper for new artists and he takes advantage of pointed and relevant data to make a better A&R choice. But music data companies continue to test the waters we all may one day sail because investors are voting for the future of computational creativity with large sums of money.
By Kyle Billings